Background and purpose: Neuropsychiatric disturbances are common and burdensome symptoms of dementias. Assessment and measurement of neuropsychiatric disturbances are indispensable to the management of patients with dementia and for clinical research. No standardized instruments have been developed for these purposes in Japan. The Neuropsychiatric Inventory (NPI), developed by Cummings et al., is a standardized, validated, and reliable tool to assess a wide range of neuropsychiatric derangements in patients with dementia, and has been used in clinical studies in North America. In this study, we prepared a Japanese version of the NPI and tested its reliability and usefulness in dementia patients.
Subjects and methods: The subjects were 174 patients with dementia, including Alzheimer's disease (140), frontotemporal dementia (12), vascular dementia (19), and normal pressure hydrocephalus (3), whose diagnosis was made on the basis of the results of extensive examinations including MRI and PET/SPECT of the brain and appropriate clinical diagnostic criteria for each disorder. The subjects consisted of 133 women and 41 men; their mean age was 72.6 +/- 7.7 years (S.D.; range: 49 to 88) years. An examiner (neurologist) used the NPI to interview a caregiver familiar with the patient's everyday behavior. In 24 randomly selected patients interview was repeated by another examiner (psychiatrist) one month later, and test-retest reliability was evaluated. The Clinical Dementia Rating (CDR), the Mini-Mental State Examination (MMSE), and the Alzheimer's Disease Assessment Scale (ADAS) were used to assess the severity of the dementia and cognitive impairment.
Results: The MMSE scores ranged from 5 to 29 (mean +/- SD = 19.0 +/- 4.8) and the ADAS scores ranged from 7 to 63 (24.3 +/- 10.9). The CDR score was 0.5 in 17 patients, 1 in 95 patients, 2 in 52 patients, and 3 in 10 patient. Test-retest reliability of each score was acceptably high, and the correlation coefficient for each score obtained by repeated assessments in this study was similar to the coefficient in the original report. Apathy was the most common behavior in the ten abnormal behavioral domains, and euphoria was the least common. The prevalence of abnormal findings was also comparable to the results in the original report. The total NPI scores and many of the NPI subscale scores were significantly correlated with both the severity of dementia and cognitive impairment.
Conclusions: This Japanese version of NPI, whose reliability and competency are comparable to those of the original version, is a reliable and useful tool for measuring neuropsychiatric disturbances in dementia patients.